From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Date: Fri, 6 Mar 2015 15:29:32 +0100 From: Pablo Neira Ayuso Message-ID: <20150306142932.GA15926@salvia> References: <1424705163-13428-1-git-send-email-imrep.amz@gmail.com> <20150223160619.GF24297@breakpoint.cc> <54EEF32D.2010202@amazon.de> <20150226.113431.238255529591339000.davem@davemloft.net> <54F982B5.90108@gmail.com> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <54F982B5.90108@gmail.com> Subject: Re: [Bridge] [RFC PATCH v2] bridge: make it possible for packets to traverse the bridge without hitting netfilter List-Id: Linux Ethernet Bridging List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: Imre Palik Cc: nbd@openwrt.org, imrep@amazon.de, netdev@vger.kernel.org, bridge@lists.linux-foundation.org, fw@strlen.de, linux-kernel@vger.kernel.org, aliguori@amazon.com, David Miller On Fri, Mar 06, 2015 at 11:34:29AM +0100, Imre Palik wrote: > On 02/26/15 17:34, David Miller wrote: > > From: Imre Palik > > Date: Thu, 26 Feb 2015 11:19:25 +0100 > > > >> If you are looking for peculiarities in my setup then here they are: > >> I am on 4k pages, and perf is not working :-( > >> (I am trying to fix those too, but that is far from being a low hanging fruit.) > >> So my guess would be that the packet pipeline doesn't fit in the cache/tlb > > > > Pure specualtion until you can actually use perf to measure these > > things. > > > > And I don't want to apply patches which were designed based upon > > pure speculation. > > > > I did performance measurements in the following way: > > Removed those pieces of the packet pipeline that I don't necessarily > need one-by-one. Then measured their effect on small packet > performance. > > This was the only part that produced considerable effect. > > The pure speculation was about why the effect is more than 15% > increase in packet throughput, although the code path avoided > contains way less code than 15% of the packet pipeline. It seems, > Felix Fietkau profiled similar changes, and found my guess well > founded. > > Now could anybody explain me what else is wrong with my patch? We have to come up with a more generic solution for this. These sysfs tweaks you're proposing look to me like an obscure way to tune this. From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1753911AbbCFO0A (ORCPT ); Fri, 6 Mar 2015 09:26:00 -0500 Received: from mail.us.es ([193.147.175.20]:53547 "EHLO mail.us.es" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751195AbbCFOZ6 (ORCPT ); Fri, 6 Mar 2015 09:25:58 -0500 X-Qmail-Scanner-Diagnostics: from 127.0.0.1 by antivirus6 (envelope-from , uid 501) with qmail-scanner-2.10 (clamdscan: 0.98.6/20155. spamassassin: 3.4.0. Clear:RC:1(127.0.0.1):SA:0(-103.2/7.5):. Processed in 1.967485 secs); 06 Mar 2015 14:25:56 -0000 X-Spam-ASN: AS12715 87.216.0.0/16 X-Envelope-From: pneira@us.es Date: Fri, 6 Mar 2015 15:29:32 +0100 From: Pablo Neira Ayuso To: Imre Palik Cc: David Miller , imrep@amazon.de, fw@strlen.de, bridge@lists.linux-foundation.org, stephen@networkplumber.org, netdev@vger.kernel.org, linux-kernel@vger.kernel.org, aliguori@amazon.com, nbd@openwrt.org Subject: Re: [RFC PATCH v2] bridge: make it possible for packets to traverse the bridge without hitting netfilter Message-ID: <20150306142932.GA15926@salvia> References: <1424705163-13428-1-git-send-email-imrep.amz@gmail.com> <20150223160619.GF24297@breakpoint.cc> <54EEF32D.2010202@amazon.de> <20150226.113431.238255529591339000.davem@davemloft.net> <54F982B5.90108@gmail.com> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <54F982B5.90108@gmail.com> User-Agent: Mutt/1.5.21 (2010-09-15) Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On Fri, Mar 06, 2015 at 11:34:29AM +0100, Imre Palik wrote: > On 02/26/15 17:34, David Miller wrote: > > From: Imre Palik > > Date: Thu, 26 Feb 2015 11:19:25 +0100 > > > >> If you are looking for peculiarities in my setup then here they are: > >> I am on 4k pages, and perf is not working :-( > >> (I am trying to fix those too, but that is far from being a low hanging fruit.) > >> So my guess would be that the packet pipeline doesn't fit in the cache/tlb > > > > Pure specualtion until you can actually use perf to measure these > > things. > > > > And I don't want to apply patches which were designed based upon > > pure speculation. > > > > I did performance measurements in the following way: > > Removed those pieces of the packet pipeline that I don't necessarily > need one-by-one. Then measured their effect on small packet > performance. > > This was the only part that produced considerable effect. > > The pure speculation was about why the effect is more than 15% > increase in packet throughput, although the code path avoided > contains way less code than 15% of the packet pipeline. It seems, > Felix Fietkau profiled similar changes, and found my guess well > founded. > > Now could anybody explain me what else is wrong with my patch? We have to come up with a more generic solution for this. These sysfs tweaks you're proposing look to me like an obscure way to tune this.